Insight · Plateer AI Labs

Article series

LLM Inside

The core ideas you need to make sense of an LLM, taken one at a time.

4 articles

The GS Certification Diary

Three weeks of reading 80 defect reports and rebuilding the code, the errors, and the manual

5 articles

The Harness Diary

Designing the execution environment an AI agent works in, from the ground up

10 articles

The Ontology Diary

Starting from the questions, building a knowledge graph, then working on quality and search

13 articles

All articles

Tech NoteThe Harness Diary2026.07.01

The tools were listed but couldn't be called (Part 9)

An allowlist meant to narrow search scope also dropped the tools the user had connected by hand. Name, schema, and permission at three moments.

김진수

Tech NoteThe Ontology Diary2026.06.30

The job reported success and the graph was empty (Part 9)

Switching to a small local model left batches unfinished, and jobs recorded as successful left an empty graph. Truncation arrived as empty results.

김진수

Tech NoteThe Ontology Diary2026.06.24

Document dedup was merging structured tables (Part 8)

We built ten CSV tables and the post-processed graph held four classes. A rule for tidying document concepts had merged schema identifiers.

김진수

Field Report2026.06.23

AI agents in retail: from reading documents to carrying out the work

A discussion with H, a retail and media company, on adopting AI agents — connecting document OCR to usable business data, why product-code generation is a reasoning problem, when fine-tuning is not required, and how to carry a PoC into production.

PPlateer AI Labs

Tech NoteThe Harness Diary2026.06.22

Letting the nearest memory always win was wrong (Part 8)

A session request is more specific than a platform policy, so it seemed it should win. Left that way, a user could override policy.

김진수

InsightInsight from our research and the field

The tools our agent built for itself were not canvas nodes